English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Existing deep learning models separate JPEG artifacts suppression from thedecoding protocol as independent task. In this work, we take one step forwardto design a true end-to-end heterogeneous residual convolutional neural network(HR-CNN) with spectrum decomposition and heterogeneous reconstructionmechanism. Benefitting from the full CNN architecture and GPU acceleration, theproposed model considerably improves the reconstruction efficiency. Numericalexperiments show that the overall reconstruction speed reaches to the samemagnitude of the standard CPU JPEG decoding protocol, while both decoding andartifacts suppression are completed together. We formulate the JPEG artifactssuppression task as an interactive process of decoding and image detailreconstructions. A heterogeneous, fully convolutional, mechanism is proposed toparticularly address the uncorrelated nature of different spectral channels.Directly starting from the JPEG code in k-space, the network first extracts thespectral samples channel by channel, and restores the spectral snapshots withexpanded throughput. These intermediate snapshots are then heterogeneouslydecoded and merged into the pixel space image. A cascaded residual learningsegment is designed to further enhance the image details. Experiments verifythat the model achieves outstanding performance in JPEG artifacts suppression,while its full convolutional operations and elegant network structure offershigher computational efficiency for practical online usage compared with otherdeep learning models on this topic.

End-to-End JPEG Decoding and Artifacts Suppression Using Heterogeneous Residual Convolutional Neural Network